Research On Imaging Spot Detection Method In Confocal Microscopy System

With the requirement of human elements on the microstructure morphology detection accuracy is higher and higher,confocal microscopy imaging method because of its special three-dimensional tomographic imaging force,non-contact measurement and protecting the object surface,can real-timely detect living cells without the advantages of the special processing,and is widely used for material development,the cell test and analysis,the tiny components quality inspection.With the development of precision machining technology,the traditional confocal microscopy detection method is difficult to meet the requirements of the high precision detection for the surface topography of components with a certain radius of curvature and inclination angle.In this paper,a confocal microscopic imaging method based on virtual pinhole is proposed.By dynamically tracking and detecting the spot position,the influence of the angle change of the measured surface on the axial characteristics of the system is eliminated for effectively improving the axial resolution of the system.The main tasks of the project"Research on Spot Tracking Detection Method in Confocal Microscopic System"include:(1)The three-dimensional coherent imaging model of the confocal system under the tilt of the measured surface is established.Based on the principle of confocal microscopy and the theory of diffraction optics,the model describes the axial response characteristics of the system when the surface of the sample is tilted.The influence of the tilt of the measured surface on the spot position and light intensity is given quantitatively by simulation calculation and it is verified that the axial resolution of the confocal microscopy system will be reduced when the tilt of the measured surface changes.The establishment of this model also lays a solid theoretical foundation for the follow-up research.(2)A fast multi-objective connected domain extraction algorithm based on region growth is proposed.In order to solve the problem of dual-path spot tracking detection in the split pupil shift focus confocal system,the image is processed by bilateral nonlinear filtering and OTUS segmentation method and the detection region of dual-path confocal spot is extracted by region growth method.By this method,the processing area of spot location is reduced and the image processing speed of the system is improved effectively.(3)An improved least square spot center localization algorithm is proposed.In order to obtain the virtual pinhole center in real time,through the comparison of several spot center location algorithms,the iterative least square circle fitting approximation method is selected.Compared with the Hough algorithm and center of mass methods,the deviation of center coordinate can be reduced by 33.47%and the running time can be reduced by 31.85%at least which can meet the needs of fast real-time location.In addition,the method has strong anti-interference ability and can track the center of the confocal spot stably.Finally,an experimental platform is built to verify the three-dimensional imaging model of the confocal system under the tilt of the measured surface and the accuracy of the spot tracking and positioning algorithm of the confocal system is verified.In addition,a virtual pinhole is used to complete the system characteristic experiment of the confocal system for the measurement of the inclined surface.The experimental results show that when the measurement objective NA₁=0.4 the tilted surface with the maximum inclination of 12°can be measured by using the spot tracking detection method.The measurement range of the system can reach 13?m and the axial resolution is 10nm.Compared with the hard pinhole the axial resolution can be improved by 20%.